Subversion Repositories Kolibri OS

Rev

Go to most recent revision | Blame | Last modification | View Log | RSS feed

  1.  
  2.                             Mesa 5.1 release notes
  3.  
  4.                               December 17, 2003
  5.  
  6.                                 PLEASE READ!!!!
  7.  
  8.  
  9.  
  10. Introduction
  11. ------------
  12.  
  13. Mesa uses an even/odd version number scheme like the Linux kernel.
  14. Even-numbered versions (such as 5.0) designate stable releases.
  15. Odd-numbered versions (such as 5.1) designate new developmental releases.
  16.  
  17.  
  18. Bug fixes
  19. ---------
  20. See the VERSIONS file for a list of bugs fixed in this release.
  21.  
  22.  
  23.  
  24. New Features in Mesa 5.1
  25. ------------------------
  26.  
  27. GL_ARB_vertex_program / GL_ARB_fragment_program
  28.    Michal Krol and Karl Rasche implemented these extensions.  Thanks!
  29.    Be aware that there may be some rough edges and lurking bugs.
  30.  
  31. GL_ATI_texture_env_combine3 extension
  32.    This adds a few new texture combine modes.
  33.    Contributed by Ian Romanick.
  34.  
  35. GL_SGI_texture_color_table
  36.    Adds a color table lookup to the RGBA texture path.  There's a separate
  37.    color table for each texture unit.
  38.    Contributed by Eric Plante.
  39.  
  40. GL_NV_fragment_program
  41.    NVIDIA's fragment-level programming feature.
  42.    Possible lurking bugs:
  43.      - the DDX and DDY commands aren't fully tested
  44.      - there may be bugs in the parser
  45.      - the TEX and TXP instructions both do perspective correction
  46.      - the pack/unpack instructions may not be correct
  47.  
  48. GL_EXT_depth_bounds_test
  49.    This extension adds a scissor-like test for the Z axis.  It's used to
  50.    optimize stencil-volume shadow algorithms.
  51.  
  52. GL_NV_light_max_exponent
  53.    Lifts the 128 limit for max light exponent.
  54.  
  55. GL_EXT_texture_rectangle
  56.    Identical to GL_NV_texture_rectangle
  57.  
  58. GL_ARB_occlusion_query
  59.    Useful for visibility-based culling.
  60.  
  61. GL_ARB_texture_non_power_of_two
  62.    Removes the restriction that texture dimensions must be powers of two.
  63.  
  64. GL_ARB_vertex_buffer_object
  65.    Allows server-side vertex arrays, optimized host/card data transfers, etc.
  66.  
  67. GL_ARB_point_sprite
  68.    ARB-approved version of GL_NV_point_sprite.  Basically allows textures
  69.    to be applied to points.
  70.  
  71. GL_IBM_multimode_draw_arrays
  72.    Allows multiple vertex arrays to be drawn with one call, including arrays
  73.    of different types of primitives.
  74.  
  75. GL_SUN_multi_draw_arrays
  76.    An alias for GL_EXT_multi_draw_arrays, standard in OpenGL 1.4.
  77.  
  78. Faster glDrawPixels / glCopyPixels in X11 driver
  79.    If your X screen is 32bpp, glDrawPixels to the front color buffer will
  80.    be accelerated (via XPutImage()) if the image format is GL_BGRA and the
  81.    type is GL_UNSIGNED_BYTE.  No raster operations, such as depth test,
  82.    blend, fog, etc. can be enabled.
  83.  
  84.    If your X screen is 16bpp, glDrawPixels to the front color buffer will
  85.    be accelerated (via XPutImage()) if the image format is GL_RGB and the
  86.    type is GL_UNSIGNED_SHORT_5_6_5.  No raster operations, such as depth
  87.    test, blend, fog, etc. can be enabled.
  88.  
  89.    glCopyPixels() calls for the front color buffer will be accelerated
  90.    (via XCopyArea()) if no raster operations, such as depth test, blend,
  91.    fog, pixel zoom, etc. are enabled.
  92.  
  93.    The speed-up over typical software rendering is a factor of 10 for
  94.    glDrawPixels and 100 for glCopyPixels.
  95.  
  96.  
  97. With the addition of GL_ARB_occlusion_query, GL_ARB_vertex_buffer_object,
  98. GL_ARB_texture_non_power_of_two and GL_EXT_shadow_funcs, Mesa 5.1 supports
  99. all the new features of OpenGL 1.5.  Mesa 6.0 (the next stable release)
  100. will advertise GL_VERSION = "1.5".
  101.  
  102.  
  103.  
  104. Vertex/Fragment program debugger
  105. --------------------------------
  106.  
  107. GL_MESA_program_debug is an experimental extension to support
  108. interactive debugging of vertex and fragment programs.  See the
  109. docs/specs/OLD/MESA_program_debug.spec file for details.
  110.  
  111. The bulk of the vertex/fragment program debugger is implemented
  112. outside of Mesa.  The GL_MESA_program_debug extension just has minimal
  113. hooks for stopping running programs and inspecting programs.
  114.  
  115. The progs/tests/debugger.c (only in CVS) program is an example of how
  116. the extension can be used.  Presently, the debugger code and demo code
  117. is in the same file.  Eventually the debugger code should be moved
  118. into a reusable module.
  119.  
  120. As it is now, the demo lets you set breakpoings in vertex/fragment
  121. programs, single step, and print intermediate register values.  It's
  122. basically just a proof of concept.
  123.  
  124.  
  125.  
  126. Directory tree reorganization
  127. -----------------------------
  128.  
  129. The directory structure for Mesa has been overhauled to improve its layout.
  130. All source code for Mesa, GLU, GLUT, etc is now under the src/ directory
  131. in appropriate subdirectories.
  132.  
  133. The Mesa source code and drivers has been reorganized under src/mesa/.
  134.  
  135. All demonstration programs and tests are now in subdirectories under progs/.
  136.  
  137.  
  138.  
  139. Build System Changes
  140. --------------------
  141.  
  142. The GNU automake/autoconf support has been removed.  As it was, it seldom
  143. worked on anything but Linux.  The Mesa developers aren't big fans of
  144. automake/autoconf/libtool and didn't have the time to maintain it.
  145. If someone wants to contribute new automake/autoconf support (and is
  146. willing to maintain it), it may be re-incorporated into Mesa, subject
  147. to some requirements.
  148.  
  149. The "old style" makefile system has been updated:
  150.    1. Make-config has been trimmed down to fewer, modern configurations.
  151.    2. Most of the bin/mklib.* scripts have been rolled into a new "mklib"
  152.       script that works on all sorts of systems.  There are probably some
  153.       bugs in it, but it's been tested on Linux, SunOS 5.8 and IRIX 6.5.
  154.       Improvements/contributes are greatly appreciated.
  155.    3. The Makefile.X11 files have been cleaned up in various ways
  156.  
  157.  
  158.  
  159. Source File Changes
  160. -------------------
  161.  
  162. The mmath.[ch] files are obsolete.  Their contents have been moved
  163. into the imports.[ch] and macros.[ch] files.
  164.  
  165. The files related to vertex and fragment programming have changed.
  166. Old files:
  167.         vpexec.[ch]
  168.         vpparse.[ch]
  169.         vpstate.[ch]
  170. New files:
  171.         program.[ch]              - generic ARB/NV program code
  172.         arbprogram.[ch]           - ARB program API functions
  173.         arbfragparse.[ch]         - ARB fragment program parsing
  174.         arbvertparse.[ch]         - ARB vertex program parsing
  175.         arbparse.[ch]             - ARB vertex/fragment parsing
  176.         arbparse_syn.h            - vertex/fragment program syntax
  177.         nvprogram.[ch]            - NV program API functions
  178.         nvvertprog.h              - NV vertex program definitions
  179.         nvfragprog.h              - NV fragment program definitions
  180.         nvvertparse.[ch]          - NV vertex program parser
  181.         nvfragparse.[ch]          - NV fragment program parser
  182.         nvvertexec.[ch]           - NV vertex program execution
  183.         swrast/s_nvfragprog.[ch]  - NV fragment program execution
  184.  
  185. The files related to per-vertex handling have changed.
  186. Old files:
  187.         tnl/t_eval_api.c          - old per-vertex code
  188.         tnl/t_imm_alloc.c         - old per-vertex code
  189.         tnl/t_imm_api.c           - old per-vertex code
  190.         tnl/t_imm_debug.c         - old per-vertex code
  191.         tnl/t_imm_dlist.c         - old per-vertex code
  192.         tnl/t_imm_elt.c           - old per-vertex code
  193.         tnl/t_imm_eval.c          - old per-vertex code
  194.         tnl/t_imm_exec.c          - old per-vertex code
  195.         tnl/t_imm_fixup.c         - old per-vertex code
  196.         tnl/t_vtx_sse.c           - old per-vertex code
  197.         tnl/t_vtx_x86.c           - old per-vertex code
  198. New files:
  199.         tnl/t_save_api.c          - new per-vertex code
  200.         tnl/t_save_loopback.c     - new per-vertex code
  201.         tnl/t_save_playback.c     - new per-vertex code
  202.         tnl/t_vtx_eval.c          - old per-vertex code
  203.  
  204. Other new files:
  205.         bufferobj.[ch]          - GL_ARB_vertex_buffer_object functions
  206.         version.h               - defines the Mesa version info
  207.  
  208. Other removed files:
  209.         swrast/s_histogram.[ch]   - moved into src/histogram.c
  210.  
  211.  
  212.  
  213. Other Changes
  214. -------------
  215.  
  216. The ctx->Driver.CreateTexture function has been removed - it wasn't used.
  217.  
  218. New device driver hook functions:
  219.    NewTextureObject   - used to allocate struct gl_texture_objects
  220.    NewTextureImage    - used to allocate struct gl_texture_images
  221.  
  222. New ctx->Texture._EnabledCoordUnits field:
  223.    With the addition of GL_NV_fragment_program we may need to interpolate
  224.    various sets of texture coordinates even when the corresponding texture
  225.    unit is not enabled.  That is, glEnable(GL_TEXTURE_xD) may never get
  226.    called but we still may have to interpolate texture coordinates across
  227.    triangles so that the fragment program will get them.
  228.    This new field indicates which sets of texture coordinates are needed.
  229.    If a bit is set in the ctx->Texture._EnabledUnits bitmask is set, the
  230.    same bit MUST be set in ctx->Texture._EnabledCoordUnits.
  231.  
  232. The ctx->_TriangleCaps field is deprecated.
  233.    Instead of testing the DD_* bits in _TriangleCaps, you should instead
  234.    directly test the relevant state variables, or use one of the helper
  235.    functions like NEED_SECONDARY_COLOR() at the bottom of context.h
  236.    While testing _TriangleCaps bits was fast, it was kludgey, and setting
  237.    the bits in the first place could be error prone.
  238.  
  239. New vertex processing code.
  240.    The code behind glBegin, glEnd, glVertex, glNormal, etc. has been
  241.    totally rewritten.  It's a cleaner implementation now and should use
  242.    less memory. (Keith)
  243.  
  244.  
  245.  
  246. To Do
  247. -----
  248. Add screen-awareness to fakeglx.c
  249.  
  250.  
  251.  
  252.  
  253. Device Driver Status
  254. --------------------
  255.  
  256. A number of Mesa's software drivers haven't been actively maintained for
  257. some time.  We rely on volunteers to maintain many of these drivers.
  258. Here's the current status of all included drivers:
  259.  
  260.  
  261. Driver                  Status
  262. ----------------------  ---------------------
  263. XMesa (Xlib)            implements OpenGL 1.4
  264. OSMesa (off-screen)     implements OpenGL 1.4
  265. FX (3dfx Voodoo1/2)     implements OpenGL 1.3
  266. SVGA                    implements OpenGL 1.3
  267. Wind River UGL          implements OpenGL 1.3
  268. Windows/Win32           implements OpenGL 1.4
  269. DJGPP                   implements OpenGL 1.4
  270. GGI                     implements OpenGL 1.3
  271. BeOS                    implements OpenGL 1.4
  272. Allegro                 needs updating
  273. D3D                     needs updating
  274.  
  275. Note: supporting OpenGL 1.4 (vs. 1.3 or 1.2) usually only requires that the
  276. driver call the _mesa_enable_1_4_extensions() function.
  277.  
  278.  
  279. ----------------------------------------------------------------------
  280.